DOCSEARCHbyAlgolia
Submit Docs

SVAHNAR AI Developer Documentation

website·Indexed 31 days ago

https://docs.svahnar.com

Developer documentation for SVAHNAR AI, covering the AI Framework, getting started guides, and tools for integrating AI with CRM, email, and other systems.

Tokens
956
Snippets
0
Records
10
Agent score
0%

What's inside SVAHNAR AI

  1. Overview of SVAHNAR AI Agent Platform

    SVAHNAR is an AI agent platform designed for building, deploying, and managing intelligent agents at scale. It supports multi-agent orchestration through 'agentic networks,' allowing for complex, collaborative AI workflows.

  2. Convert SVAHNAR agent configurations from JSON to YAML

    SVAHNAR provides a guide for converting agent configurations from JSON format to YAML format, which is required for use within the SVAHNAR platform.

  3. Implement Tools and Human-in-the-Loop (HITL) in SVAHNAR

    SVAHNAR agents use tools to perform real-world actions. Key tool-related features include:

    • Custom Instructions: Tool-specific guidance to refine agent behavior during individual tool calls.
    • Human-in-the-Loop (HITL): A mechanism allowing human administrators to approve, modify, or reject AI agent actions before they are executed.

  4. Configure Knowledge Repository for RAG

    SVAHNAR's KnowledgeRepository enables Retrieval-Augmented Generation (RAG) by allowing AI agents to perform semantic searches over indexed data. You can connect various data sources to the KnowledgeRepository using the following connectors:

    • AWS S3: Connect S3 buckets as a knowledge source.
    • Confluence: Integrate Atlassian Confluence pages.
    • SharePoint: Connect Microsoft SharePoint sites.
    • URL: Use public web URLs as a knowledge source.

  5. Manage the SVAHNAR AI Agent Lifecycle

    Developers can manage agents through the following operations:

    • Creation: Deploy agents via the console or YAML configuration.
    • Validation: Validate YAML configurations before deployment to catch errors.
    • Testing: Verify behavior, responses, and tool usage before going live.
    • Execution: Run agents, trigger workflows, and monitor real-time responses.
    • Maintenance: Retrieve agent details, list all agents, update configurations/settings, or permanently delete agents.

  6. Extend Agent Capabilities with Model Context Protocol (MCP)

    SVAHNAR supports the Model Context Protocol (MCP), allowing developers to configure and use MCP servers to provide agents with dynamic tool integration.

  7. Set up OAuth 2.0 connections for external services

    SVAHNAR supports secure OAuth 2.0 connections to integrate with external services.

    • Standard Connections: Built-in support for Google, Slack, Microsoft Outlook, Atlassian, Box, GitHub, Notion, Salesforce, and SharePoint.
    • Custom OAuth Clients: Developers can register their own OAuth 2.0 applications within SVAHNAR to connect any service, including private MCP (Model Context Protocol) servers, using their own client credentials.

  8. Configure SVAHNAR AI Agents and Agentic Networks

    SVAHNAR agents are configured using YAML. The configuration framework includes several key components:

    • Initialization: Settings for agent startup.
    • LLM Settings: Model selection (supporting GPT-4o, GPT-5, Llama, and Deepseek), temperature, and other parameters.
    • Tools: Assignment of capabilities to agents for real-world actions.
    • Agent Functions: Logic to control behavior, prompts, and decision-making.
    • Edges: Routing logic and communication paths between agents in a multi-agent network.

  9. Integrate external tools with SVAHNAR AI agents

    SVAHNAR provides a wide array of pre-built tool integrations that allow AI agents to interact with external services. These include:

    • Custom APIs & Logic: REST API (custom HTTP requests), AWS Lambda (serverless functions).
    • Communication: Gmail, Outlook, Slack, Discord, Twilio SMS, Twilio WhatsApp, AWS SES (email).
    • Productivity & Management: Google Calendar, Google Drive, Notion, Calendly, Jira, GitHub, GitLab.
    • CRM & Business: HubSpot, Salesforce.
    • Research & Data: Arxiv, Google Scholar, Wikipedia, Wikidata, WolframAlpha, Tavily (real-time web search), Webpage Search (URL extraction).
    • Finance: Google Finance, Yahoo Finance, Zerodha Kite.
    • Specialized Data: Pinecone (vector search), Knowledge Repo (semantic search for RAG), Database Tool (SQL read/write), Database Tool Read-Only.
    • Automation: Zapier.

  10. Secure Secret Management with SVAHNAR Key Vault

    SVAHNAR Key Vault provides centralized secret management with encryption at rest. It is used to securely store, manage, and revoke API keys and credentials required by AI agents.